Introduction
Beginning in 2024, a stream of “incredible” growth metrics from VC-backed startups in Silicon Valley began to show up, especially in Artificial Intelligence (AI) startups. Their reported Annual Recurring Revenues (ARRs) showed almost unbelievable growth.
But how reliable is this most amazing of growth metrics?
Take, for example, the case of Midjourney, an AI service that creates images from text descriptions, or “prompts”. It is a generative AI tool that operates on the chat platform Discord and allows users to type a command and receive AI-generated art in return. In less than three years, its Annual Recurring Revenue (ARR) grew from zero to $200 million.
Another example is ElevenLabs, a voice AI startup that develops AI software that generates realistic and expressive speech from text, with tools for voice cloning, a voice library, and creating custom voices from text descriptions. In 20 months, it saw its ARR soar from zero to near $100 million.
There are so many other examples. Lovable, an AI-powered platform that allows users to create websites and full-stack web applications by describing their ideas in natural language prompts, went from zero to $17 million in ARR in just three months. Decagon, an enterprise AI platform that helps support teams automate customer service using large language models (LLMs), achieved “seven figures” in ARR within its first six months. It claims that its AI agents can autonomously handle tasks like answering product questions, processing refunds, and cancelling subscriptions — helping businesses reduce support costs and scale faster. Cursor AI, is an AI-powered code editor that assists developers by using large language models to help write, debug, and refactor code using natural language prompts. The company’s ARR went from zero to $100 million in a year.
These numbers are eye-popping. But are they reliable? And how does the ARR impact investor interest and company valuations?
This article explores what appears to be the new frontier of creative accounting.
What is Annual Recurring Revenue (ARR)?
ARR is a key financial metric used primarily by subscription-based businesses to measure the total value of recurring revenue components in a business’s revenue stream, normalised on an annual basis. It is a crucial indicator for Software as a Service (SaaS) companies and other subscription-based businesses, as it provides insight into the company’s revenue health and growth potential (Nansubuga and Kowalkowski, 2024).
How is the ARR calculated?
The key variable in the calculation is Recurring Revenue. This could be via (1) monthly or yearly subscriptions that customers pay for continued access to a service and via (2) contracts, i.e., long-term agreements that ensure a steady revenue stream over a specified period.
The main difference is that ARR (Annual Recurring Revenue) is a forward-looking metric that represents the current, predictable revenue from active subscriptions, while CARR (Contracted Annual Recurring Revenue) is an even more forward-looking metric that includes committed, but not yet billed, future revenue from signed contracts, including anticipated expansions. In essence, ARR shows the present revenue snapshot, while CARR forecasts future potential and growth based on existing contracts.
There are some Exclusions in both calculations. Both ARR and CARR do not include one-time fees like setup fees, consulting fees, or any other non-recurring charges. Also excluded are variable charges, such as usage-based charges, which fluctuate and are not predictable.
For AI-powered, usage-based models, Token usage is the key component used for calculating the ARR, because it directly translates to the revenue generated by usage-based pricing models.
The ARR is calculated by taking the Monthly Recurring Revenue (MRR) and multiplying it by 12. This provides a straightforward view of the expected revenue over a year.
ARR = MRR x 12
Alternatively, if the business bills annually, ARR can be calculated by summing up all the annual subscription fees.
It is as simple as that, as shown in the following example:
Example: A SaaS company has a mixed billing system. It has 1000 customers, each paying $50 in subscriptions on a monthly plan, and 2000 customers on an annual plan, paying $400 per year.
MRR from monthly subscriptions = 1000 customers x $50 = $50,000
Therefore, ARR from monthly subscriptions = $50,000 x 12 = $600,000
ARR from annual subscriptions = 2000 customers x $400 = $800,000
Therefore, Total ARR = $600,000 + $800,000 = $1,400,000
Many companies offer multiple subscription tiers, such as Basic, Pro, and Enterprise, and the ARR from each tier must be summed to calculate the total company ARR.
Token Usage
To incorporate token usage into ARR, companies must determine the revenue from token usage by first determining the cost per token. Then it must estimate customer usage by analysing user behaviour to predict how many tokens a customer will use over a given period, e.g., monthly. Then it can calculate monthly usage revenue by multiplying the number of tokens by the cost per token to calculate the revenue for that month. Finally, to obtain the annualised usage revenue, multiply the monthly usage revenue by 12 to obtain the annual usage revenue.
How Does the ARR impact Company Valuation?
Annual Recurring Revenue (ARR) significantly impacts company valuation by indicating a company’s revenue predictability, scalability, and customer retention, which are all key factors for investors. Higher ARR suggests a more stable and predictable revenue stream, leading to a higher valuation, especially for SaaS and subscription-based businesses, which are often valued based on a multiple of their ARR. A strong ARR multiple reflects a healthy business with sustainable competitive advantages (SaaS Capital, 2025).
Investors, especially VC investors, often use the ARR multiple to determine how much they are willing to pay for each dollar of a company’s recurring revenue.
Strong ARR companies often command significantly higher valuations, sometimes surpassing those relying on perpetual licences by six times. For example, Adobe’s stock price nearly tripled after it shifted to a subscription model, demonstrating the market’s preference for this stability (Nurkka, et. al., 2017).
In its purest form, the ARR should provide a clear, predictable revenue stream, reducing uncertainty compared to one-time sales. This financial stability is highly attractive to investors and allows for more accurate forecasting of future earnings Consistent ARR growth indicates that a company can scale its operations efficiently and grow its customer base. High ARR, particularly when looking at metrics like retention and churn, is considered a strong indicator of customer satisfaction and loyalty, which is crucial for long-term success.
Valuations of Start-ups.
Unlike public companies, start-up companies (especially Tech start-ups) are not monitored the way that public companies are Whilst public companies have to report quarterly. to a country’s Securities and Investment Commission, VC investors do not necessarily look for audit reports of the start-up companies they invest in. Whilst there is a financial due diligence process that may involve an informal audit before a VC invests, it is more likely a game played with trust.
And in the SaaS era (which technically started in the 1990s and gained steam through the 2000s), trust in ARR came comparatively easily. In the earlier period, there was an agreed-upon set of conventions. For example, annual per-seat pricing was standard, where one user pays for one year, and then accounts expanded by adding multiple users. There was a clear separation between ARR and CARR (signed contract value before activation) and recognised revenue (actual revenue booked). Typically, 80% to 90% of CARR would convert to ARR, and you could predictably chart a company’s expansion, relying on low churn rates and steady customers.
There were, in short, standardised methods of calculating ARR.
How ARR Creativity Became the Favoured Metric in AI Valuations
As outlined earlier, until about the early 2020s, everyone agreed with these terms in the SaaS world. All this has now changed. Anna Barber, a partner at VC firm M13 now says:
“The numbers then were a lot harder to manipulate, because people had a general understanding of what things had to mean. Today, we don’t know what things have to mean in the same way. So, there is a lot of confusion and, maybe, obfuscation.” (Garfinkle, 2025).
Although absolute revenue clarity was not necessarily always there in the SaaS era, as the cloud computing wave started to take shape, ARR started to get a little erratic. People began to question the suitability of subscription revenue as a proxy for ARR. However, it was the emergence of AI that created a whole new layer of uncertainty.
Nnamdi Okike, cofounder and managing partner of 645 Ventures, says:
“Investors wanted to keep evaluating companies as SaaS-predictable, so they tried to shoehorn those elements into ‘recurring’ revenue. It doesn’t truly work, but it worked well enough for investors to keep doing it. Now AI has shown up with a whole new set of elements, and it would be better for investors to finally create new metrics to represent this new reality.” (Garfinkle, 2025).
Creative Revenue Recognition Across Industries
Revenue itself, on a fundamental level, features both core truths and discretionary realities. As such, whilst there was a significant amount of variation across industries, there were also widely accepted optimal accounting principles pertaining to revenue recognition. Until proven otherwise, adopting an optimistic perspective on revenue is not illegal, and many even consider it a long-standing tradition. But that can still cause problems (or crises) down the line.
Traditionally, companies book revenue when the service is provided and/or when the goods are delivered. Depending on how the contracts are written, depending on how clear those stated objectives or benchmarks are noted, and/or just the industry in general, there is some room for discretion and perhaps misreporting, intentional or not. Also, there are often incentives tied to revenue for management and members of the sales teams.
In general terms, however, there are normal red flags around revenue that accounting experts watch for. A primary concern among external auditors is to watch out for revenue that is being creatively inflated. There are always many things auditors are looking for, but a company potentially trying to manipulate the revenue numbers to achieve a goal that really contradicts objective reporting is always a red flag.
Pressure in the Tech Industry for High ARRs
Today, there is significant pressure to be the company that went from zero to $100 million in ARR in the shortest number of days.
As discussed earlier, the ARR metric came to be a favourite of VCs and startups through the software-as-a-service (SaaS) wave starting in the 2000s, when it was widely accepted as a trusted proxy for a stable startup with a reliable source of revenue and a reasonably shored-up future.
However, as billions flowed across the venture capital ecosystem into AI startups, some mere months old, the vaunted, trusted ARR metric has morphed into something much harder to recognise. There is now a massive amount of pressure on AI-focused founders to report significant ARR growth at earlier stages than ever before: If they are not generating revenue immediately, it raises the question of what their current focus is. Founders—to keep up with the Joneses—are counting all sorts of things as “long-term revenue” that are, to be blunt, nothing your regular accounting professor would recognise as legitimate.
Exacerbating the pressure is the fact that more VCs than ever are trying to funnel capital into possible winners at a time when there is no certainty about what evaluating success or traction even looks like. Throughout the ’90s, VC as an industry grew to more than 700 firms managing about $143 billion. Today, there are more than 3,000 VC firms, according to the National Venture Capital Association, managing more than $360 billion, with some projections suggesting venture will be a more than $700 billion industry by 2029 (Ceppos, 2025).
Creative accounting has a long history of emerging during economic booms, a practice that dates back to the Gilded Age when inflating assets, understating liabilities, and bribery were common. More recently, the dotcom boom and the lead-up to the Great Recession revealed practices such as ‘channel stuffing’, ‘roundtripping revenue’, and the infamous ‘special purpose entities’ in creatively reporting revenue. Now, industry watchers are beginning to raise concerns about ARR.
The problem is that so much of this is essentially ‘speculative’ revenue. This situation is different from Google signing a data centre contract with a provider. That is real future revenue. In most cases, a startup temporarily uses another company’s name, product, or service while claiming to have a contractual relationship. This certainly is not recurring revenue.
Creative Accounting in ARR Calculations
Today, ARR is in what could be described as an awkward phase, where there are some AI startups that are trying to use the metric with sincerity, but their business dynamics are just too different from traditional SaaS businesses. Many are still in an experimentation phase, trying all sorts of products on short-term pilots, creating high churn risk.
Furthermore, AI services often have unpredictable token usage, which refers to the amount of text that AI processes to understand language. (More tokens equal more usage, and more complicated queries require more token usage, by extension.) So, a few “inference whales” like OpenAI and Anthropic have massive pricing power and can skew costs, making AI startups’ financial structures fundamentally different from traditional SaaS businesses.
The result? Founders are counting pilots, one-time deals, or inactivated contracts as recurring revenue. For example, some startups are claiming “booked ARR”—numbers based on what customers might pay in the future rather than what they actually are paying now—even though contracts frequently have provisions that let customers opt out at any time for any reason.
Companies are entering into contracts that include kill provisions, which allow them to claim booked ARR while providing their customers with a way out. Surely it should concern VC investors if a startup company books a million-dollar-a-year contract, includes it in their ARR calculation, only to have it cancelled for no reason within three months?
A case in point is Cluely, a startup born of controversy after Founder Roy Lee posted in a viral X thread saying he was suspended by Columbia University because he and a co-founder developed a tool to cheat on job interviews for software engineers. He claimed that he turned around and created a product and startup out of the tech, originally using the marketing tagline that it helps you “cheat on everything.” However, the startup’s controversial history has not stopped businesses from showing interest in Cluely’s product. Now, due to big-league VCs backing it, Cluely has toned down its marketing to “Everything You Need. Before You Ask.” Cluely now claims to have doubled ARR to $7 million over a week by signing a public company that doubled its annual contract with them to $2.5 million. However, Lee declined to name the company (Temkin, 2025).
The Circular Startup Ecosystem
There are also broader sociological changes making the ARR shenanigans possible. A part of the fault lies in well-established accelerators which have standardised “what to say” to raise money. This approach has encouraged metric manipulation.
Further, many of these startups ultimately sell to other startups circuitously, making things even more insular. More than private equity, more than even banking, venture capital has an ‘in’ crowd, i.e., a certain sort of person gets funded with a certain sort of business model.
Conclusion
The emphasis on ARR is ultimately reflective of a wider reckoning in ventures overall. Not only are there more VCs (and more capital) than ever, but priorities are in flux. Ilya Strebulaev, a professor at the Stanford Graduate School of Business and coauthor of The Venture Mindset, says:
“Generally, historically, there’s been an important trade-off in the venture capital industry between profitability and growth. But roiled by geopolitical tensions and macroeconomic uncertainty, that pendulum has been changing over time. I think venture capitalists are now spending more effort on profitability today than in the past and are spending more effort on revenue. But that doesn’t mean the trade-off between profitability and growth has evaporated—absolutely not.” (Garfinkle, 2025).
The consensus among VCs seems to be that ARR will ultimately not be the way forward at all: Priya Saiprasad, general partner at Touring Capital, says:
“The classic SaaS model is dying as we speak. We shouldn’t be using classic SaaS terms to measure these companies; we shouldn’t be using the language of it. So, we should all, collectively as an industry, evolve to a new set of metrics we feel comfortable measuring these companies by.” (Garfinkle, 2025).
Ultimately, smart investors will develop new ways to assess AI businesses, focusing on retention, daily active usage, and unit economics. If a bubble bursts, VCs and founders who inflated ARR may suffer most. This is an equity-driven boom, and the main losers in equity-driven booms like the one right now are the ones who made the bets.
References:
Ceppos, Robin (2025), NVCA Releases 2025 Yearbook Showcasing 2024 VC Trends, National Venture Capital Association, March 27. https://nvca.org/press_releases/nvca-releases-2025-yearbook-showcasing-2024-vc-trends/
Garfinkle, Allie (2025), “Founders are using creative accounting to boost lofty ‘ARR’—the hottest startup metric in Silicon Valley”, Fortune, Sept 28. https://fortune.com/2025/09/28/how-is-arr-calculated-startups-venture/
Nansubuga, Brenda and Kowalkowski, Christian (2024), “Moving to subscriptions: service growth through business model innovation in consumer and business markets”, Journal of
Service Management, 35(6):185-215
Nurkka, Jaakko; Waltl, Josef and Alexy, Oliver (2017), “How Investors React When Companies Announce They’re Moving to a SaaS Business Model”, Harvard Business Review, January 13, https://hbr.org/2017/01/how-investors-react-when-companies-announce-theyre-moving-to-a-saas-business-model
SaaS Capital (2025), “SaaS Valuation Multiples: Understanding the New Normal”, SaaS Capital Blog Posts, August 6, https://www.saas-capital.com/blog-posts/saas-valuation-multiples-understanding-the-new-normal/
Temkin, Marina (2025), “Cluely’s ARR doubled in a week to $7M, founder Roy Lee says. But rivals are coming”, Techcrunch, July 3. https://techcrunch.com/2025/07/03/cluelys-arr-doubled-in-a-week-to-7m-founder-roy-lee-says-but-rivals-are-coming/